Stock market evidence on the international transmission channels of US monetary policy surprises - Tim D. Maurer, Thomas Nitschka
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
Stock market evidence on the international transmission channels of US monetary policy surprises Tim D. Maurer, Thomas Nitschka SNB Working Papers 10/2020
Legal Issues DISCLAIMER The views expressed in this paper are those of the author(s) and do not necessarily represent those of the Swiss National Bank. Working Papers describe research in progress. Their aim is to elicit comments and to further debate. COPYRIGHT© The Swiss National Bank (SNB) respects all third-party rights, in particular rights relating to works protected by copyright (infor- mation or data, wordings and depictions, to the extent that these are of an individual character). SNB publications containing a reference to a copyright (© Swiss National Bank/SNB, Zurich/year, or similar) may, under copyright law, only be used (reproduced, used via the internet, etc.) for non-commercial purposes and provided that the source is menti- oned. Their use for commercial purposes is only permitted with the prior express consent of the SNB. General information and data published without reference to a copyright may be used without mentioning the source. To the extent that the information and data clearly derive from outside sources, the users of such information and data are obliged to respect any existing copyrights and to obtain the right of use from the relevant outside source themselves. LIMITATION OF LIABILITY The SNB accepts no responsibility for any information it provides. Under no circumstances will it accept any liability for losses or damage which may result from the use of such information. This limitation of liability applies, in particular, to the topicality, accuracy, validity and availability of the information. ISSN 1660-7716 (printed version) ISSN 1660-7724 (online version) © 2020 by Swiss National Bank, Börsenstrasse 15, P.O. Box, CH-8022 Zurich
Stock market evidence on the international
transmission channels of US monetary policy
surprises∗
Tim D. Maurer†, Thomas Nitschka‡
23 June 2020
Abstract
We decompose unexpected movements in the stock market returns
of 40 countries into different news components to assess why expan-
sionary US monetary policy surprises are good news for stock markets.
Our results suggest that prior to the zero lower bound (ZLB) period,
federal funds rate surprises affect foreign stock markets mainly be-
cause such surprises are associated with news about future real interest
rates. The effects of forward guidance surprises are negligible. At the
∗
We are grateful to Eric Swanson for sharing high-frequency data on US monetary
policy surprises and to Roman Elbel for help with implementing bootstrap procedures.
This paper benefitted from comments and suggestions by an anonymous referee of the
SNB Working Paper Series, Petra Gerlach, Mico Loretan, Enno Mammen, Aleksander
Welfe (discussant) and participants of the SNB Brown Bag Seminar, an internal research
seminar of Danmarks Nationalbank as well as participants of the 2019 Paris Financial
Management Conference and the 2019 Meeting of Swiss Economists Abroad. Parts of
this research have been conducted while Tim D. Maurer was at the Swiss National Bank.
Any errors and omissions are our own. The views, opinions, findings, and conclusions
or recommendations expressed in this paper are strictly those of the authors. They do
not necessarily reflect the views of the Swiss National Bank or Danmarks Nationalbank.
Neither the Swiss National Bank nor Danmarks Nationalbank take responsibility for any
errors or omissions in, or for the correctness of, the information contained in this paper.
†
tima.eco@cbs.dk, Copenhagen Business School and Danmarks Nationalbank
‡
thomas.nitschka@snb.ch, Swiss National BankZLB, large-scale asset purchases (LSAP) reflect more than commit-
ment to forward guidance. LSAP surprises constitute cash-flow news,
while unanticipated forward guidance primarily reflects real interest
rate news.
JEL: E44, E52, F36, G15
KEYWORDS: international spillovers, news, monetary policy, stock
returns, vector autoregression
1 Introduction
Expansionary US monetary policy surprises tend to be good news for stock
markets worldwide (Ammer et al., 2010; Brusa et al., 2019; Ehrmann and
Fratzscher, 2009; Jiang et al., 2019; Miranda-Aggripino and Rey, 2018; Thor-
becke, 1997; Wongswan, 2009). While this evidence is persuasive, the un-
derlying reasons for foreign stock markets’ responses to US monetary policy
surprises remain elusive.
In a textbook world, monetary policymakers set the short-term inter-
est rate, influence expectations about future short-term interest rates and
thus affect asset prices. However, there is growing evidence, based on high-
frequency data, that non-monetary information, e.g., the central bank’s growth
outlook or risk assessments, is an important part of central bank communica-
tion to which financial markets respond (Campbell et al., 2012; Campbell et
al., 2017; Cieslak and Schrimpf, 2019; Jarocinski and Karadi, 2019; Kroencke
et al., 2019; Nakamura and Steinsson, 2018; Miranda-Aggripino and Ricco,
2019).1 Indeed, Bernanke and Kuttner (2005) show that federal funds rate
surprises affect US stock markets to a large extent through their impact on
expectations of future excess returns.
This background raises the question of the channels through which US
1
Madeira and Madeira (2019) show in a high-frequency event study that even the
degree of consent among members of the Fed’s Federal Open Market Committee affects
stock prices.
1
1monetary policy surprises influence foreign stock markets. Do foreign stock
markets respond to US monetary policy surprises because US monetary pol-
icy affects expectations of real interest rates? Do they respond because the
Fed influences expectations about future cash flows or about future risk pre-
mia? Do unconventional monetary policy measures, e.g., large-scale asset
purchases (LSAP), affect stock markets through the same channels as con-
ventional measures, such as changes in the federal funds rate?
To shed light on these issues, we build on Bernanke and Kuttner (2005)
and extend their analysis by assessing whether US monetary policy surprises
occurring in the period from 1991 to 2015 are associated with unexpected
stock return movements in 40 developed and emerging economies.2 Moreover,
we combine the vector autoregressive model (VAR) of Bernanke and Kuttner
(2005) with the decomposition of unexpected variation in excess returns on
US dollar (USD)-denominated stock indices performed by Ammer and Mei
(1996). In addition to components reflecting cash-flow news, real interest
rate news and news about future excess returns, this decomposition gives us
a return component related to real exchange rate news. Hence, we are able to
directly evaluate whether the different types of US monetary policy surprises
affect expectations about foreign countries’ real exchange rates against the
USD. Finally, since our sample period covers the zero lower bound (ZLB)
period for the fed funds rate, we also take forward guidance and LSAP sur-
prises into account. This allows us to assess whether we observe differences
in the impact on foreign stock markets among the three different types of US
monetary policy surprises. We employ the proxies of fed funds rate, forward
guidance and LSAP surprises proposed by Swanson (2018) in our empirical
2
We use the MSCI classification of developed and emerging markets as of December
2018 for this distinction and use MSCI standard stock market indices in our empirical
analysis. Note that inclusion of a firm’s stock in the MSCI indices requires the stock to
fulfil certain quality criteria and to offer the opportunity for both domestic and foreign
investors to invest in it. Therefore, using the MSCI indices allows us to directly compare
evidence for developed markets with that for emerging markets.
2
2analysis.3
Distinguishing between the different components of unexpected foreign
stock market returns reveals that fed funds rate surprises statistically signifi-
cantly affect real interest rate news prior to the ZLB period. This is the main
channel through which US monetary policy surprises influenced all foreign
stock markets from 1991 to 2008. Economically, however, cash-flow and ex-
cess return news are by far more important than real interest rate news, but
the link between cash-flow and excess return news and US monetary policy
surprises is less significant in statistical terms. Forward guidance surprises
are only associated with unexpected movements in some of the foreign stock
market returns under study. In addition, there is no clear evidence of a par-
ticular channel through which forward guidance transmits to foreign stock
markets.
In the ZLB period from 2009 to 2015, forward guidance primarily affected
foreign stock markets through real interest rate news. We find an associa-
tion between forward guidance surprises and cash-flow and expected return
news for some of the countries under study as well, but the overall statis-
tical significance for these news components is weaker than for real interest
rate news. The signs of the responses also vary across countries. Further-
more, our findings suggest that LSAP surprises have an impact on foreign
stock markets that goes beyond commitment to forward guidance (Swanson,
2018). LSAP surprises mainly affect foreign stock markets through revisions
in expectations about the future stream of dividends (cash-flow news), which
are related to the long-term growth prospects of an economy (Ammer and
Mei, 1996; Campbell, 1991; Campbell and Vuolteenaho, 2004; Lucas, 1978).
LSAP surprises have no impact on real interest rate news.
Moreover, US monetary policy surprises are only weakly associated with
revisions in expectations about future real exchange rates. We find that
3
We are grateful to Eric Swanson for sharing the monetary policy surprise series with
us.
3
3LSAP surprises affect stock market returns of emerging markets through
real exchange rate news. However, real exchange rate news only explains a
small proportion of the unexpected variation in stock market returns. Thus,
our findings are in line with the recent event study evidence by Brusa et al.
(2019) and suggest that the impact of US monetary policy surprises on foreign
stock markets through real exchange rate news is very small. However, this
does not mean that the overall exchange rate effects of US monetary policy
surprises are negligible. Our results are focused on unexpected movements
in stock market returns and not on macroeconomic aggregates.
Overall, our empirical assessments highlight that the link between mon-
etary policy surprises in the US and unexpected stock return movements
abroad is significant for developed and emerging markets and for economies
with fixed or floating exchange rate systems. The channels, i.e., return
news components related to monetary policy surprises, differ between the
period before the ZLB and the ZLB period but are similar for both emerging
and developed markets. Our analysis, however, provides only little evidence
in favour of the recently proposed information effect of Fed communication
(Nakamura and Steinsson, 2018). According to the information effect, Fed
communication of tightening (loosening) monetary policy constitutes good
(bad) news about the economic outlook. We do not find compelling evi-
dence of this effect in our analysis. While the direction of the link between
forward guidance surprises and unexpected stock market movements is not
clear for some of the countries in our sample, expansionary fed funds rate or
LSAP surprises constitute good news for foreign stock markets in the vast
majority of cases. Bauer and Swanson (2020) argue that such stock market
reactions are inconsistent with an information effect of the Fed’s monetary
policy communication.
It is beyond the scope of this paper to pinpoint the exact economic and
theoretical mechanism underlying our empirical findings. However, we view
our results as consistent with an explanation based on the existence of a
4
4global financial cycle. The global financial cycle, i.e., common movements in
risky asset prices, cross-border borrowing and lending, and macroeconomic
aggregates, appears to be affected by changes in monetary policy in the
US, the dominant country in the international financial system (Rey, 2016;
Miranda-Aggripino and Rey, 2018). Jiang et al. (2019) rationalize the obser-
vation of a global financial cycle with a model in which US monetary policy
affects the supply of safe USD-denominated assets. These assets play a cen-
tral role as collateral for banks worldwide and thus open the door for US
monetary policy to influence the risk-bearing capacity of banks across the
globe and thus their pricing of risk as reflected in stock prices and returns.
The remainder of the paper is organized as follows. Section 2 highlights
the main contributions of our paper in the context of the related literature.
Section 3 provides the background for the decomposition of unexpected stock
market returns into different news components. Section 4 introduces the
empirical framework to obtain the effects of monetary policy surprises on the
news components. Details on the data are outlined in section 5. Section 6
presents the main empirical results, and section 7 concludes. The appendix
at the end of the paper provides further details.
2 Main contributions and related literature
The empirical analyses of our paper rely on the VAR models of Campbell
(1991) and Ammer and Mei (1996) to decompose unexpected movements in
asset returns into components reflecting revisions in expectations about cash
flows, expected returns, real interest rates and the real exchange rate. Distin-
guishing between different news components is important because cash-flow
news reflects a revision in expectations about the entire stream of dividends
and is thus related to the long-term growth prospects of an economy (Lu-
cas, 1978). Cash-flow news hence has a persistent impact on asset returns.
Expected return news has a temporary impact because a capital loss due to
5
5adverse expected return news today will be compensated by the expectation
of higher returns in the future (Campbell and Vuolteenaho, 2004).
To link the different return news components to monetary policy sur-
prises, we employ a VAR with exogenous variables that is similar to the
empirical framework of Bernanke and Kuttner (2005). However, we do not
focus on the US but analyse a sample of 40 developed and emerging markets.
Furthermore, we evaluate the responses of different stock return news com-
ponents to three different types of monetary policy surprises (fed funds rate,
forward guidance, and LSAP) because our sample includes the ZLB period
during which the Fed resorted to unconventional measures. We thus com-
plement Rogers et al. (2018), who use a VAR framework with exogenously
approximated monetary policy surprises related to the different facets of the
Fed’s monetary policy to study the impact of US monetary policy surprises on
bond premia and forward exchange rate premia (relative to USD) in the euro
area, Japan and the UK. We focus on stock markets and employ a similar,
but slightly different, VAR framework that allows for a detailed distinction
between different international transmission channels of US monetary policy,
and our sample includes more countries.
Moreover, Paul (2019) uses monetary policy surprises as exogenous vari-
ables in a VAR system that includes macroeconomic aggregates as well as US
stock and house prices. He shows that fed funds rate surprises are directly in-
terpretable as monetary policy shocks while other monetary policy surprises,
e.g., forward guidance, cannot be taken as a direct proxy of monetary policy
shocks. This is why, for example, Gertler and Karadi (2015) use monetary
policy surprises as an external instrument to identify monetary policy shocks.
In this context, it is important to note that we do not attempt to look at
monetary policy shocks. We focus on surprises that are, by construction,
orthogonal to each other, such that we can assess differences in the responses
of the stock return news components to the different surprises. Furthermore,
we do not look explicitly at macroeconomic aggregates in our VAR setting,
6
6but we note that the different stock market return news components are in-
terpretable in a macroeconomic way. For example, Ammer and Mei (1996)
analyse the cross-country correlations of cash-flow and expected return news
between the US and the UK to evaluate the degree of real and financial in-
tegration between the two countries. They argue that cash-flow news can
be interpreted as news about the real economy, i.e., economic growth, while
news about future returns reflects news about the price of risk.
In addition, our paper complements event studies of the effects of US
monetary policy on foreign economies’ bond and stock markets as well as
the responses of foreign exchange rates to US monetary policy surprises
(Albagli et al., 2019; Ammer et al., 2010; Ehrmann and Fratzscher, 2009;
Hausman and Wongswan, 2011; Swanson, 2019; Wongswan, 2009). Unlike
these studies, our monthly VAR estimations allow us to distinguish the ef-
fects of monetary policy surprises on different components of asset returns.
These components reflect revisions in expectations and are thus not directly
comparable to simple changes in stock prices. However, our stock return
decompositions have to include short-term real interest rates and real ex-
change rates directly into the VAR system. Therefore, we need information
about consumer price indices. The highest frequency at which these data are
available is monthly.
Moreover, our assessment of the link between the three different types
of US monetary policy surprises and the four news components of emerg-
ing economies’ stock market returns gives us a better understanding of the
channels through which monetary policy surprises in the US affect emerg-
ing markets. This assessment complements studies that assess the impact of
US monetary policy on emerging markets’ macroeconomic aggregates (e.g.,
Vicondoa, 2019) and financial markets (e.g., Akinci, 2013; Fratzscher et al.,
2018), and it corroborates that monetary policy surprises in the US have
important real economic implications for emerging markets.
Finally, our paper contributes to and is consistent with the growing lit-
7
7erature on the emergence of a global financial cycle and its importance for
the valuation of assets. Miranda-Aggripino and Rey (2018) show that one
global risk factor accounts for non-negligible amounts of the variation in a
broad set of risky asset prices around the world. This global factor in asset
prices responds to US monetary policy changes. Jiang et al. (2019) provide
a theoretical mechanism for this empirical finding. The mechanism is based
on the notion that there is special demand for USD-denominated safe assets
from financial intermediaries worldwide (e.g., McCauley and McGuire, 2009).
US monetary policy, e.g., LSAP, affects the supply of these safe assets, such
that the risk-bearing capacity and hence the risk appetite of financial inter-
mediaries are affected. Our results of a statistically significant link between
US monetary policy surprises and unexpected variation in the stock market
returns of both developed and emerging countries are consistent with this
economic mechanism.
3 Economic background: decomposing unex-
pected variation in stock market returns
Our empirical assessments of the effects of US monetary policy surprises on
foreign stock markets are based on a decomposition of unexpected variation
in stock market returns into different components. This decomposition relies
on a dynamic accounting identity that relates asset returns to expected cash
flows and discount rates (Campbell and Shiller, 1988; Campbell, 1991). This
accounting identity is based on the two-period version of the present value
model that links current stock prices to dividends and returns, i.e.,
Pt+1 + Dt+1
1 + rt+1 = (1)
Pt
with r, denoting the net return on the stock, P, the stock price excluding
dividends, and D, dividends. Equation (1) implies that if stock prices and
8
8dividends are non-stationary, then returns should be stationary. However,
stock returns vary at business-cycle frequencies, as risk aversion varies over
the business cycle. To allow for time variation in returns, Campbell and
Shiller (1988) propose a log-linear approximation of equation (1) around the
mean dividend-price ratio. This approximation yields
rt+1 ≈ k + ρpt+1 + (1 − ρ)dt+1 − pt (2)
in which lower-case letters denote logarithms of the variables. The letter
k summarizes the constant terms that result from the Taylor expansion, and
ρ = 1/(1+exp(d−p)) is a weight that follows from the log-linearization. This
weight depends on the long-run mean of the log dividend-price ratio, d–p,
around which equation (1) is linearized. Rearranging equation (2) for the
stock price, pt ≈ k + ρpt+1 + (1 − ρ)dt+1 − rt+1 and expanding to the infinite
horizon gives a representation of unexpected variation in stock returns. This
expansion imposes the condition that discounted stock prices cannot grow
forever, lim ρj pt+j = 0.
j→∞
Taking expectations on both sides of the equation gives
∞
k
pt = + Et ρj ((1 − ρ)dt+1+j − rt+1+j ) (3)
1−ρ j=0
with E being the expectation operator conditional on information at time
t. Substituting equation (3) into equation (2), Campbell (1991) highlights
that unexpected changes in stock returns reflect either news (revisions in
expectations) of dividend growth or future discount rates, i.e.,
∞
∞
rt+1 − Et rt+1 = (Et+1 − Et ) j
ρ ∆dt+1+j − (Et+1 − Et ) ρj rt+1+j (4)
j=0 j=0
To study excess returns, i.e., stock returns in excess of a short-term debt
9
9rate, equation (4) can be rewritten as
∞ ∞ ∞
rxt+1 −Et rxt+1 = (Et+1 −Et ) ρj ∆dt+1+j − ρj rrt+1+j − ρj rxt+1+j
j=0 j=0 j=1
(5)
Here, we decompose the discount rate, r, into the short-term real interest
rates, rr, and a risk premium term, rx.
For notational convenience, equation (5) can be rewritten as
cf
T
ηt+1 = ηt+1 rr
− ηt+1 rx
− ηt+1 (6)
with ηt+1
T
≡ rxt+1 −Et rxt+1 being the unexpected stock market excess return,
ηt+1 ≡ (Et+1 − Et ) ∞
cf
j=0 ρ ∆dt+1+j , being the news of future cash flows (divi-
j
dends), ηt+1
rr
≡ (Et+1 −Et ) ∞ j=0 ρ rrt+1+j , being the news of the real interest
j
rate, and finally ηt+1
rx
≡ (Et+1 − Et ) ∞ j=1 ρ rxt+1+j being the news of future
j
excess returns, which can be interpreted as a proxy of expected risk premia.
Following from this accounting identity, a positive surprise movement in the
excess stock market return is associated with positive dividend news, lower
than expected real interest rates, lower than expected future excess returns
or an arbitrary combination of the three.
Equation (5) also holds for foreign stock returns and can be written as
∞ ∞ ∞
rx∗t+1 −Et rx∗t+1 = (Et+1 −Et ) ρj ∆d∗t+1+j − ∗
ρj rrt+1+j − ρj rx∗t+1+j
j=0 j=0 j=1
(7)
with asterisk superscripts denoting a foreign variable. Following Ammer and
Mei (1996), we focus on the foreign stock return denominated in USD in
excess of the US short-term interest rate, f rx.
The innovation in foreign excess stock returns can be expressed as
10
10
∞ ∞
∗ j
f rxt+1 − Et f rxt+1 = (Et+1 − Et ) (ρ ) ∆d∗t+1+j − (ρ∗ )j rrt+1+j
j=0 j=0
∞
∞
(8)
− (ρ∗ )j ∆qt+1+j − (ρ∗ )j rx∗t+1+j
j=0 j=1
Here, q represents the real exchange rate, which is denominated in foreign
currency as a unit of domestic currency (USD). Again for the purpose of
notational convenience, equation (8) can be rewritten as
cf q
T
ft+1 = ft+1 rr
− ft+1 − ft+1 rx
− ft+1 (9)
in which ft+1
T
≡ f rxt+1 −Et f rxt+1 denotes the unexpected foreign stock mar-
ket excess return denominated in US dollars, ft+1 cf
≡ (Et+1 −Et ) ∞ ∗ j ∗
j=0 (ρ ) ∆dt+1+i
is the news of future foreign cash flows (dividends), ft+1
rr
≡ (Et+1 −Et ) ∞ ∗ j
j=0 (ρ ) rrt+1+j
q
represents the news about the US real interest rate, ft+1 ≡ (Et+1 −Et ) ∞ ∗ j
j=0 (ρ ) ∆qt+1+i
gives the news about the exchange rate and finally ft+1rx
≡ (Et+1 −Et ) ∞ ∗ j ∗
j=1 (ρ ) rxt+1+j
denotes the news about future foreign excess returns.
The effect of f cf , f rr and f rx on unexpected excess returns can be inter-
preted analogously to their corresponding news term in equation (6). The
intuition behind the negative sign on exchange rate news f q is that, ceteris
paribus, news about a future USD appreciation must have an adverse impact
on the returns of foreign assets denominated in USD.
4 Empirical framework
4.1 The VAR
Estimates of all news terms defined in section 3 can be computed using a VAR
that includes the excess return on a stock market and variables that predict
returns. In the literature, this has been done by a VAR specification as in
11
11equation (10). For notational flexibility, this process is stated in first order.
This notation also represents a higher-order process written as a first-order
VAR in companion form
zt+1 = Γzt + t+1 (10)
Here, zt+1 is a vector of the endogenous variables, Γ denotes the companion
matrix and is an i.i.d. error vector. As we are interested in the effect
of monetary policy surprises, abbreviated with ∆Mtu , on the different news
terms, we include the surprise series as exogenous variables in the VAR
u
zt+1 = Γzt + φ∆Mt+1 + Ψt+1 , (11)
where φ captures the response of the endogenous variables in the VAR to
the contemporaneous monetary policy surprises and Ψt+1 is the new error
term. This error term is, by construction, orthogonal to our monetary policy
surprise series ∆M u . This splits our baseline error term t+1 in equation
(10) into innovations in monetary policy and innovations in all other factors
unrelated to monetary policy.
The estimated regression model has the form
p−1
A(L)yt+1 = yt+1 − Al+1 yt−l − φ∆Mt+1
u
= Ψt+1 (12)
l=0
where A(L) is a polynomial of order p and yt+1 = [y1,t+1 , y2,t+1 , . . . , yK−1,t+1 , yK,t+1 ]
is a column vector with K endogenous variables in the VAR. Thus, the com-
panion matrix is of the form
A1 A2 ··· Ap−1 Ap
I 0 ··· 0 0
0 I ··· 0 0
Γ= .
. .. .. .. ..
. . . . .
0 0 ··· I 0
12
12with I, the identity matrix, and Ai both being matrices of dimensions (K ×
K), where i ∈ {1, 2, . . . , p − 1, p}. In this case, the left-hand-side vector in
equation (10) is zt+1 = [yt+1 , yt , . . . , yt+1−p , yt−p ] . As the VAR in equation
(11) is also presented in companion form, it can be rewritten as:
yt+1 A1 A2 ··· Ap−1 Ap yt
yt I 0 ··· 0 0 yt−1
.. 0 I ··· 0 0 ..
.= .
. .. .. .. ..
yt+1−p .. . . . . yt−p
yt−p 0 0 ··· I 0 yt−p−1
Φ 0 ··· 0 u
ψt+1
∆m t+1
0 0 · · · 0 0
0
· · · 0 0
+ 0 0 . + ,
. .. .. .. .. .
. .
. . . . .
0
0 0 ··· 0 0
where Φ is a (K × s)-matrix of the monetary surprise coefficients, with s being the
number of the surprise series and φ being the corresponding
matrix in companion
Φ 0 ··· 0 0
0 0 ··· 0 0
form with dimensions ((K ∗ p) × (K ∗ s)) such that φ ≡
.
0 0 ··· 0 0.
.. .. .. ..
.. . . . .
0 0 ··· 0 0
The vector of monetary policy surprises u
∆Mt+1 ≡ [∆mut+1 , 0, 0, . . . , 0] has di-
mensions ((3∗p)×1), and the error vector Ψt+1 ≡ [ψt+1 , 0, 0, . . . , 0] has dimensions
((K ∗ p) × 1).
In the subsequence, we report the results from estimating equation (11) for each
of the two sample periods: from 1991 to 2008 (the pre-ZLB period) and from 2009
to 2015 (the ZLB period). Alternatively, one could apply a two-step procedure.
Bernanke and Kuttner (2005) argue that vector ∆Mt+1
u represents a prediction
error from a rational forecast at time t and should therefore be orthogonal to
the endogenous variables zt . Thus, one can obtain consistent estimates of both Γ
and φ from a two-step estimation. The first step is to run the baseline VAR in
equation (10) without the exogenous monetary policy surprises to estimate Γ and
to extract t+1 . To obtain an estimate of φ, one has to regress the VAR’s one-
13
13month-ahead forecast error t+1 in a second step on the monetary policy surprises
u . Applying the two-step procedure leads to qualitatively similar results as
∆Mt+1
our preferred direct, one-stage estimation of equation (11).
4.2 The news
To estimate the domestic (US) news terms η T ,η cf ,η rr and η rx , we stack four endoge-
nous variables into the VAR. The first variable in the VAR is the US excess return
rxt+1 , the second variable is the US real interest rate rrt+1 , the third variable is
the US yield spread yst+1 and finally the fourth variable is the dividend-price ratio
δt+1 .4 In addition, we include three different exogenous monetary policy surprise se-
ries ∆Mt+1
u = [∆iu , ∆g u , ∆lu ] . These series are based on the high-frequency
t+1 t+1 t+1
identification proposed by Swanson (2018), with ∆iut representing unanticipated
changes in the fed funds rate, ∆gtu representing unanticipated changes in forward
guidance and ∆lu representing unanticipated changes in large-scale asset purchases.
5 With t+1 ≡ (φ∆Mt+1 t+1 ) and using Γ and φ, one can then calculate the
u +Ψ
different domestic news terms as follows
T
ηt+1 = λ1 t+1 (13)
rx
ηt+1 = λ1 ρΓ(I − ρΓ)−1 t+1 (14)
rr
ηt+1 = λ2 (I − ρΓ)−1 t+1 (15)
cf T rx rr
ηt+1 = ηt+1 + ηt+1 + ηt+1 , (16)
where ρ is a discount factor. In addition, we define λ1 = [1, 0, 0, 0] to pick rx and
λ2 in an analogous fashion.
Based on the approach proposed by Ammer and Mei (1996), we stack seven
endogenous variables into the VAR for foreign stock market returns to estimate for-
eign news terms f T ,f cf ,f rr , f rx and f q . In addition to the four variables used to es-
4
The data section presents the exact definitions of the variables.
5
Appendix A briefly describes the high-frequency identification strategy and section 5
describes how we convert the high-frequency surprises on event days into a monthly time
series for the empirical analysis.
14
14timate domestic news terms, we include the foreign stock market excess retun rx∗t+1
, the change in the log real exchange rate ∆qt+1 and the foreign dividend-price ratio
∗ . The variables are ordered such that z = [rx
t+1 , rxt+1 , rrt+1 , ∆qt+1 , δt+1 , yst+1 , δt+1 ] .
δt+1 ∗ ∗
t
Again using Γ and φ, one can then calculate the different foreign news terms as
follows
T
ft+1 = Λ2 t+1 (17)
rx
ft+1 = Λ2 ρ∗ Γ(I − ρ∗ Γ)−1 t+1 (18)
rr
ft+1 = Λ3 (I − ρ∗ Γ)−1 t+1 (19)
q
ft+1 = Λ4 (1 − ρ∗ )(I − ρ∗ Γ)−1 t+1 (20)
cf T rx rr q
ft+1 = ft+1 + ft+1 + ft+1 + ft+1 (21)
Here, we define Λ2 = [0, 1, 0, 0, 0, 0, 0] to pick rx∗ and define Λ3 and Λ4 in an
analogous fashion.
We need rx∗t+1 ,rrt+1 and ∆qt+1 to estimate the different news components.
All of the other variables should be predictors of foreign stock market returns.
The choice of variables in our VAR system for all of the foreign (non-US) stock
markets reflects our aim of basing the empirical results for developed and emerging
markets on similar VAR systems to make the results comparable. That is why, for
example, we do not include foreign markets’ yield spreads into the system, because
data on government bond yields for a sufficiently long time span are not available
for most of the emerging markets in our sample. In addition, based on the advice
of Engsted et al. (2012), we include foreign stock markets’ dividend-price ratios
in the VAR systems to address criticism related to the practice of obtaining cash-
flow news indirectly from the VAR, which possibly overstates its importance (Chen
and Zhao, 2009). Engsted et al. (2012) show that by including the dividend-price
ratio (and thus the stock price) into the VAR system, it does not matter whether
cash-flow news or discount rate news is obtained directly from the VAR estimates
because the dividend-price ratio incorporates information about the stock price
level. The stock price level has to be part of the VAR system because the return
decomposition only holds conditional on the stock price itself being a part of the
15
15information set. Moreover, we include not only the US stock market return but also
the US yield spread and the US dividend-price ratio in the VAR system because
Rapach et al. (2013) show that US stock market returns predict foreign stock
market returns both in sample and out of sample, whereas foreign stock markets
do not lead the US stock market. More generally, their results suggest that US
financial market variables observed today are informative about future foreign stock
market returns.
4.3 The effect of monetary policy surprises
To assess what news components (revisions in expectations) account for the reac-
tion of foreign stock market excess returns to the US monetary policy surprises,
we exploit that the error t+1 from our baseline VAR in equation (10) can be de-
composed into innovations in monetary policy ∆Mt+1
u and innovations related to
all possible factors other than monetary policy. This allows us to rewrite equations
(17)-(21), which define how we calculate the different news terms, as follows
T u
ft+1 = Λ2 (φ∆Mt+1 + Ψt+1 ) (22)
rx
ft+1 = Λ2 ρ∗ Γ(I − ρ∗ Γ)−1 (φ∆Mt+1
u
+ Ψt+1 ) (23)
rr
ft+1 = Λ3 (I − ρ∗ Γ)−1 (φ∆Mt+1
u
+ Ψt+1 ) (24)
q
ft+1 = Λ4 (1 − ρ∗ )(I − ρ∗ Γ)−1 (φ∆Mt+1
u
+ Ψt+1 ) (25)
cf T rx rr q
ft+1 = ft+1 + ft+1 + ft+1 + ft+1 (26)
The effect of US monetary policy surprises on the foreign excess stock market
return (e.g., total news ft+1
T ) becomes
u
Λ2 φ∆Mt+1 . (27)
The sources that account for this stock market reaction are the effects of US
monetary policy surprises on the news terms that reflect revisions about the dis-
counted sums of expected future excess returns, real interest rates, real exchange
rate changes, and cash flows. The response of excess return news to US monetary
16
16policy surprises is
Λ2 ρ∗ Γ(I − ρ∗ Γ)−1 φ∆Mt+1
u
, (28)
the response of real interest rate news is
Λ3 (I − ρ∗ Γ)−1 φ∆Mt+1
u
, (29)
the response of real exchange rate news is
Λ4 (1 − ρ∗ )(I − ρ∗ Γ)−1 φ∆Mt+1
u
, (30)
and finally, the response of cash-flow news is
Λ2 + Λ2 ρ∗ Γ(I − ρ∗ Γ)−1 + Λ3 (I − ρ∗ Γ)−1 + Λ4 (1 − ρ∗ )(I − ρ∗ Γ)−1 φ∆Mt+1
u
. (31)
To make inferences about the statistical significance of the responses of foreign
stock markets’ news components to US monetary policy surprises, we take into
account that the news terms are estimated variables and depend on the VAR
parameters. The details are presented in section 6.
5 Data
Our empirical assessments are based on data measured at monthly frequency.6 The
sample period is restricted by the availability of the US monetary policy surprise
series. These series start in July 1991 and end in October 2015. The surprises are
identified through a high-frequency analysis of various US financial market prices
in a narrow time window (30 minutes) around FOMC announcements (Gürkaynak
et al., 2005; Swanson, 2018). These surprises are measured on the day of the
FOMC announcement, and they are, by construction, orthogonal to each other. To
perform the VAR-based decompositions, we have to convert these event-day data
6
In principle, one could run the VARs at daily frequency. However, to perform our
preferred decomposition of unexpected stock market returns, one has to directly include
real interest rates and real exchange rates into the VAR system. One needs consumer
price indices (CPI) to compute these variables. The highest frequency at which CPI data
are available is monthly.
17
17into periodic time series data at the monthly frequency. To do so, we follow Romer
and Romer (2004) and Gertler and Karadi (2015) and first calculate the cumulative
daily surprise series. In a second step, we take monthly averages of these series. The
first differences (month by month) of these series are our measures of US monetary
policy surprises converted to the monthly frequency. A negative value of a surprise
series reflects surprisingly looser monetary policy in that particular dimension of
the monetary policy stance. Monetary policy is then more expansionary than
expected. The surprise series are separately estimated for the period from 1991
to 2008, distinguishing between surprises related to the level of the fed funds rate
and those related to forward guidance for the fed funds rate’s future path, and
the period from 2009 to 2015, distinguishing between surprises related to forward
guidance and those related to the LSAP. We present the results for these surprise
series in the subsequence. Alternatively, one could identify the three surprise series
over the full sample period (Swanson, 2018). The qualitative results do not depend
on this choice. Figure (1) presents the time series of the monthly surprises series,
which all have a mean of 0 and standard deviation of 1. According to Swanson
(2018), a one-standard-deviation change in the fed funds rate surprise factor can
be thought of as an 8.8-basis-point surprise change in the Fed funds rate, a one-
standard-deviation change in the forward guidance surprise factor corresponds to a
6-basis-point change in the Eurodollar futures rate, and a one-standard-deviation
change in the LSAP surprise factor roughly corresponds to a 225 billion USD
surprise LSAP announcement.7
[Figure 1 about here]
We use end-of-month values of USD-denominated MSCI stock market indices to
compute log returns on the 40 developed and emerging economies’ stock markets
in our sample. Using end-of-month values helps to ensure that the (monthly)
monetary policy surprises occur before the monthly stock return reaction. We use
the MSCI classification to distinguish between developed and emerging markets.
Furthermore, inclusion in MSCI market indices requires the stock to fulfil certain
quality criteria and the opportunity for both domestic and foreign investors to
7
For more details see appendix A.
18
18invest in it. Therefore, using the MSCI indices allows us to directly compare
evidence for developed markets with evidence for emerging markets. Table (1)
provides an overview of the economies under study and the start and end dates of
the stock market return data.
The MSCI indices are also used to compute log dividend-price ratios. We
compile the log dividend-price ratio as the log of the sum of monthly dividends
over the past year minus the log of this month’s MSCI price index. Dividend series
are obtained from the difference between the returns on the MSCI gross (i.e., total
return) index and the returns on the MSCI price index. The MSCI indices are
available on the MSCI website.
Excess returns are obtained as the difference between the log stock market re-
turn and the one-month US treasury bill rate. The US yield spread is defined as
yield on ten-year US government bonds minus the one-month treasury bill rate. US
real interest rates are calculated as the one-month treasury bill rate in t minus real-
ized consumer price inflation in month t. Real USD exchange rates are constructed
from nominal, bilateral USD exchange rates measured at the end of the month and
monthly consumer price indices. The source for the exchange rate, interest rate
and consumer prices series is the IMF’s International Financial Statistics.
[Table (1) about here]
6 Empirical results
This section presents our main results. We start with a brief description of the
outcome of the country-level VAR estimations. Then, we provide our estimates of
the responses of the different news components of the US and foreign stock markets
to US monetary policy surprises.
6.1 Brief overview of VAR estimation results
The VAR estimates for the US show that past values of the yield spread and the
dividend-price ratio forecast US stock market excess returns one month ahead.
19
19The sources of return predictability for the foreign stock markets8 vary widely.
However, US short-term real interest rates and the country’s own dividend-price
ratio are the variables that most often exhibit forecasting power for foreign stock
market returns. The country-specific VAR estimates are not reported to conserve
space, but they are readily available upon request.
Table (2) displays the adjusted R2 statistics from the return forecasting equa-
tion of the country-specific VARs. In the US, for example, depending on the period
and whether the exogenous variables are included in the VAR, between 2.0% and
8.6% of the one-month-ahead variation in stock market returns is explained by the
variables in the VAR system. This is within the range of the R2 statistics reported
by Campbell (1991) and Bernanke and Kuttner (2005). The US stock market ex-
hibits an average degree of stock return predictability compared with that of the
markets in the other countries in our sample. For some, such as Colombia and
Indonesia, the variables in the VAR explain more than 10% of one-month-ahead
variation in stock market returns while for others, e.g., Japan, there is essentially
no evidence of predictability (the R2 adjusted for the number of regressors falls
into negative territory). Overall, foreign stock market returns appear to be ap-
proximately equally predictable as US stock market returns.
[about here Table (2)]
Finally, we perform variance decompositions of the unexpected market returns.
These decompositions show that cash-flow and excess return news are the most
important components of unexpected stock market returns. Real interest rate
news is only of minor importance and real exchange rate news contributes virtually
nothing to the variation in unexpected stock market returns. Furthermore, we find
that expected return news is exceptionally important to understand variation in US
stock market returns (Campbell, 1991). However, for the vast majority of the other
countries, cash-flow news is the dominant driver of variation in the unexpected
stock market returns. This pattern is not unique to our sample of countries or the
sample period but has been documented before (e.g., Nitschka, 2010). The details
8
Standard information criteria suggest that one lag is optimal for all markets’ VARs in
the pre-ZLB period. The same number of lags is used in the ZLB period.
20
20of the variance decompositions are not reported to conserve space but are available
upon request.
6.2 US monetary policy surprises and stock market news
6.2.1 The period of “conventional” monetary policy: July 1991 to
December 2008
This section presents the empirical assessment of the link between the different
components of the unexpected stock market returns of the 40 countries under
study and US monetary policy surprises.
Tables (3) and (4) contain the estimates of the effects of US monetary policy
surprises on the different news series defined in equations (27)-(31) for every country
in the pre-ZLB period. Statistical significance is assessed using an error resampling
bootstrap procedure applied to equation (11). The appendix B provides more
details on the bootstrap procedure.
In our sample of developed markets, fed funds rate surprises affect the stock
markets of all the developed markets under study through US real interest rate
news. This finding is in line with the use of US interest rates as an approximation
of a world interest rate (Akinci, 2013; Uribe and Yue, 2006). As seen from the
fourth column of table (3), the impact of a one-standard-deviation surprise decline
in the fed funds rate varies between 0.6% and 1.2% p.a.. Using the metrics of the
size of surprises from Swanson (2018), this finding implies that a surprise lowering
of the fed funds rate by 25 basis points (bp) translates into a 1.7% to 3.4% p.a.
increase in unexpected, monthly foreign stock market returns.9
The responses of expected future cash flows or future excess returns are larger
than the real interest rate news reaction, but there is no evidence of statistical
significance. This finding explains why total stock market return news appears to
be unrelated to the fed funds rate surprises. Moreover, real exchange rate news is
9
Table (3) gives the coefficients in terms of a one-unit surprise in the Fed funds rate,
which corresponds to an 8.8bp change in the first fed funds rate futures contract. Thus, to
calculate the effect of a lowering of the fed funds rate by 25 bp, the coefficients in the first
five columns must be multiplied by 3.4. See appendix A for more details on Swanson’s
(2018) metric of the size of the monetary policy surprises.
21
21unrelated to the fed funds rate surprises as well.
Forward guidance surprises only affect the unexpected stock market returns of
eight (out of 21) developed stock markets in a statistically significant way. For those
countries, we find, again using the metric of surprises provided by Swanson (2018),
that a surprise decline in the fed funds rate by 25 bp in one year, implies an increase
in unexpected, monthly foreign stock market returns ("Total" in the right panel
of table(3)) of up to 75% p.a. (New Zealand). However, the decomposition of the
total stock market return news does not reveal the main channel(s) through which
forward guidance surprises affect foreign stock market returns. The coefficients
that measure the sensitivity of the different news components to forward guidance
surprises are statistically indistinguishable from zero.
Taken together, these estimates suggest that surprisingly accommodative US
monetary policy constitutes good news for foreign stock markets. US monetary
policy surprises primarily affect foreign stock markets through revisions of expec-
tations of future real interest rates. In statistical terms, the link with unexpected
stock market return components is stronger for fed funds rate surprises than for
forward guidance surprises.
[about here Table (3)]
In our sample of emerging markets, the responses of unexpected stock market
returns to US monetary policy surprises are broadly similar to those in their de-
veloped market counterparts. Fed funds rate surprises primarily affect emerging
markets’ stock returns through US real interest rate news. There is no exception.
As shown by the estimates presented in table (4), the responses of emerging mar-
kets’ stock returns to fed funds rate surprises through real interest rate news vary
between 0.9% (Hungary) and 1.4% (Mexico) p.a.; i.e., a surprise fed funds rate
decline of 25 bp translates into an increase of 2.6% to 4.0% p.a. in unexpected,
monthly stock market returns.
Compared with the evidence for developed markets, there is even less evidence
of forward guidance surprises affecting emerging markets’ unexpected stock re-
turns. Only a few stock market news components reflect a statistically significant
sensitivity to forward guidance surprises.
22
22In sum, our evidence highlights that fed funds rate surprises are associated with
unexpected stock market returns for all of the countries under study, irrespective of
their classification as developed or emerging market or their exchange rate regime.
Looser than expected US monetary policy is associated with higher than expected
foreign stock market returns.
[about here Table (4)]
6.2.2 “Unconventional” monetary policy surprises: January 2009
to October 2015
The responses of developed and emerging markets to US monetary policy surprises
in the period from January 2009 to October 2015 show clear differences from those
in the previous period of conventional monetary policy.
Table (5) presents the sensitivities to US forward guidance and LSAP surprises
for the different stock return news components of the developed markets under
study. These results suggest that forward guidance surprises in the ZLB period are
primarily associated with real interest rate news. We find a statistically significant
link for 17 of the 21 developed markets, but the magnitudes of the responses are by
far smaller than those of the responses to the fed funds rate surprises in the pre-ZLB
period. Moreover, the responses show the "wrong" (negative) sign for almost half of
the developed markets under study. The significant coefficients range from -0.35%
p.a. to 0.29% p.a. Occasionally, forward guidance surprises constitute expected
return news or cash-flow news. However, the signs of the responses are not uniform.
In some cases, surprisingly loose forward guidance leads to higher expected excess
returns, which means that stock prices today have to decline. In other cases, we
observe the opposite pattern. For most of the countries under study, there is no
statistically significant link at all.
By contrast, the overwhelming majority of the stock market returns of devel-
oped markets react to LSAP surprises in a statistically significant way. In these
cases, all of the unexpected stock market return responses have the expected sign.
Surprisingly loose monetary policy, e.g., in the form of more asset purchases than
expected, is associated with an increase in unexpected foreign stock market returns.
23
23This finding applies to 16 of the 21 developed economies in our sample and the
significant responses to a surprise one-standard-deviation change in LSAP range
from 17% p.a. to 58% p.a. LSAP surprises influence foreign stock markets through
cash-flow news, i.e., news about the real economy (Ammer and Mei, 1996). This
finding supports Swanson (2018), who argues that LSAP represents more than just
the Fed’s commitment to its forward guidance (Woodford, 2012). This result also
means that LSAP surprises have persistent effects on foreign stock market returns.
They influence revisions in expectations about the entire future stream of dividends
(Campbell and Vuolteenaho, 2004).
[about here Table (5)]
We observe similar patterns in our sample of emerging markets (see table (6)).
Forward guidance surprises are associated with real interest rate news, and, for
some countries, cash-flow news. The sizes of the real interest rate news responses
to forward guidance surprises are comparable to those of the responses reported
for developed markets. LSAP surprises also affect emerging markets’ stock returns
through revisions in expectations about future cash flows, i.e., the real economy.
This is the case for 16 out of the 19 emerging markets under study. The sensitivities
of the return news to LSAP surprises have the expected sign (surprisingly more
accommodative LSAP is associated with higher unexpected returns). There is
only one exception. Finally, we find a statistically significant association between
LSAP surprises and emerging stock markets’ real exchange rate news. However,
revisions in expectations about future real exchange rates contribute little to the
overall variation in unexpected stock market returns. This does not mean that the
impact of US monetary policy on the real exchange rates of emerging markets is
unimportant per se. It is simply relatively small in the context of the stock market
return decomposition presented in this paper.
[about here Table (6)]
In sum, these results suggest that developed and emerging stock markets re-
sponded similarly to US monetary policy surprises in the ZLB period. Our findings
also suggest that LSAP surprises not only signal commitment to forward guidance
24
24but also capture different information that is relevant for stock markets. Moreover,
the sensitivities of unexpected stock market return components to US monetary
policy surprises do not provide compelling evidence for the information effect in Fed
communication (Nakamura and Steinsson, 2018). Most of our estimates suggest
that surprisingly accommodative US monetary policy has been associated with
good news for foreign stock markets – a feature of the data that appears to be
inconsistent with the information effect (Bauer and Swanson, 2020).
7 Conclusions
This paper has used a VAR-based decomposition of unexpected variation in the
stock market returns of developed and emerging markets into different news com-
ponents to assess why foreign stock markets react to US monetary policy surprises.
Our main results highlight that the transmission of US monetary policy sur-
prises to foreign stock markets works through different channels. Prior to the ZLB
period, US monetary policy surprises in the form of surprise changes in the fed
funds rate affected developed and emerging markets primarily by influencing ex-
pectations about future real interest rates. However, we find that at the ZLB,
forward guidance and LSAP surprises affected foreign stock markets through dif-
ferent channels. This latter finding highlights that LSAP cannot be viewed only
as a signal of commitment to forward guidance.
The main findings of this paper are in line with theory and evidence of a global
financial cycle, that is influenced by US monetary policy and affects risky asset
prices worldwide. This observation raises the question of whether the effects of
US monetary policy surprises on foreign stock markets are stronger or weaker than
local monetary policy surprises. Answering this question is beyond the scope of
this paper, but shedding light on this issue could help to assess constraints on
domestic monetary policy due to the global financial cycle. This question is of
particular interest for the emerging markets under study and a fruitful avenue for
future research.
25
25References
[1] Akinci, Ö. (2013). Global financial conditions, country spreads and macroeco-
nomic fluctuations in emerging countries. Journal of International Economics
91, 358-371.
[2] Albagli, E., Ceballos, L., Claro, S., Romero, D. (2019). Channels of US mon-
etary policy spillovers to international bond markets. Journal of Financial
Economics 134, 447-473.
[3] Ammer, J., Mei, J. (1996). Measuring International Economic Linkages with
Stock Market Data. Journal of Finance 51, 1743-1763.
[4] Ammer, J. Vega, C., Wongswan, J. (2010). International Transmission of
US Monetary Policy Shocks: Evidence from Stock Prices. Journal of Money,
Credit and Banking 42, 179-198.
[5] Bauer, M., Swanson, E.T. (2020). The Fed’s response to Economic News Ex-
plains the "Fed Information Effect". Unpublished Working Paper.
[6] Bernanke, B.S., Kuttner, K.N. (2005). What Explains the Stock Market’s
Reaction to Federal Reserve Policy? Journal of Finance 60, 1221-1257.
[7] Brusa, F., Savor, P., Wilson, M. (2019). One Central Bank to Rule Them All.
forthcoming Review of Finance.
[8] Campbell, J.R., Evans, C.L., Fisher, J.D.M., Justiniano, A. (2012). Macroe-
conomic Effects of Federal Reserve Forward Guidance. Brookings Papers on
Economic Activity, Spring 2012.
[9] Campbell, J.R., Fisher, J.D.M., Justiniano, A., Melosi, L. (2017). Forward
Guidance and Macroeconomic Outcomes Since the Global Financial Crisis.
NBER Macroeconomics Annual 2016, 31, 283-357. Eds. Martin Eichenbaum
and Jonathan A. Parker, NBER, Chicago University Press.
[10] Campbell, J.Y. (1991). A variance decomposition for stock returns. Economic
Journal 101, 157-179.
26
26[11] Campbell, J.Y., Shiller, R.J. (1988). The dividend price ratio and expecta-
tion of future dividends and discount factors. Review of Financial Studies 1,
195-227.
[12] Campbell, J.Y., Vuolteenaho, T. (2004). Bad beta, good beta. American Eco-
nomic Review 94, 1249-1275.
[13] Chen, L., Zhao, X. (2009). Return Decomposition. Review of Financial Studies
22, 5213-5249.
[14] Cieslak, A., Schrimpf, A. (2019). Non-Monetary News in Central Bank Com-
munication. Journal of International Economics 118, 293-315.
[15] Ehrmann, M., Fratzscher, M. (2009). Global financial transmission of mone-
tary policy shocks. Oxford Bulletin of Economics and Statistics 71, 739-759.
[16] Engsted, T., Pedersen, T.Q., Tanggaard, C. (2012). Pitfalls in VAR based
return decompositions: A clarification. Journal of Banking and Finance 36,
1255-1265.
[17] Fratzscher, M., Lo Duca, M., Straub, R. (2018). On the International
Spillovers of US Quantitative Easing. Economic Journal 128, 330-377.
[18] Gertler, M., Karadi, P. (2015). Monetary Policy Surprises, Credit Costs, and
Economic Activity. American Economic Journal: Macroeconomics 7, 44-76.
[19] Gürkaynak, R.S., Sack, B.P., Swanson, E.T. (2005). Do Actions Speal Louder
than Words? The Response of Asset Prices to Monetary Policy Actions and
Statements. International Journal of Central Banking 1, 55-93.
[20] Hausman, J., Wongswan, J. (2011). Global asset prices and FOMC announce-
ments. Journal of International Money and Finance 30, 547-571.
[21] Jarocinski, M., Karadi, P. (2019). Deconstructing monetary policy surprise:
the role of information shocks. forthcoming American Economic Journal:
Macroeconomics.
27
27You can also read
